Review on Multiple Classifier System in Pattern Recognition

Recently many researchers concentrate on Multiple Classifier System (MCS) in pattern recognition. Pattern recognition system build in three steps i.e. database, feature extraction and classifier. MCS is Architect by combining more than one classifier i.e. either same or different classifiers for different pattern recognition applications such as emotion recognition, character recognition, face recognition etc. This review paper discuss on the designing multiple classifier system is according to topological structure i.e. series or parallel. In MCS classifiers are combined by combiner/fusion rule such as voting schemes, rank based method, Bayes approach, probability schemes etc. from study it observed that MCS improves accuracy as compared to the output of any individual classifier.